agentleFS
thewyattbrocatothewyattbrocato/design-and-copy-skills/skills/sales-conversations/SKILL.md

sales-conversations

Skill2 starsChanged 12 days agoMITView on GitHub

Use whenever a live sales conversation is being planned, run or reviewed, in any words or language - "prep me for a discovery call", "what questions should I ask this prospect", "plan a demo for this buyer", "why did this deal stall", "they said send me more information", "write a cold-call or voicemail script", "how do I handle this objection", "should I run a free trial", "critique my call notes or my demo plan". Covers the goal of the call, discovery and qualifying, sizing the problem in the buyer's own units, demo and trial or pilot plans, stalled deals and indecision, objections and price in conversation, calls and voicemail, and a dated next step. Not for written outreach sequences (email-and-sequences), whether a persuasion lever is fair (fair-persuasion), whether a claim is true (honest-claims), research meant to learn rather than sell (user-research-basics), slide craft (presentation-design) or setting prices (pricing-and-packaging).

1 other project uses a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 3 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈3

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    All if installed; if a neighbor is missing, use ordinary judgment.

    line 31
  2. Distinctive:

    If a call plan, script or demo already has a goal, a gate, buyer-owned numbers and a dated next step, say so and change at most one thing.

    line 36
  3. Very specific:

    No fact the user supplied is asked as a question.

    line 62
  4. Distinctive:

    No demo, trial or proposal appears before the gate is met.

    line 64

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Sales conversations

A sales conversation moves when the seller knows what this call is for, learns the buyer's problem in the buyer's own numbers, and ends with a next step someone owns. The usual failures are no stated goal, generic questions asked before reading what is already known, a demo before anyone has said what it must prove, an invented return figure, a stuck buyer given more fear or more questions, and a call that ends on "let me know".

Do not produce

  • A call, demo or trial plan with no goal for this conversation, or with two.
  • Questions that ask the buyer for facts the user already supplied, or a question list where no question says what number it should produce.
  • A savings, return, cost-of-problem, win-rate, customer or reference figure the user did not give, stated as fact.
  • A demo, trial, pilot or proposal offered before a felt problem, a signer and a route to purchase are on the table.
  • "Let me know", "I'll follow up", "circle back" or "check in" as the end of a call or message: the next step has no date and no owner.
  • The cost of waiting, a deadline or "the price goes up" aimed at a buyer who has already said they want to change.
  • A discount or concession in reply to a first price reaction.
  • A pressure script: a seller-made deadline, a fake approval check, guilt, a push after a no, pulling a trial away, a pilot that turns into a sale unless the buyer objects, a voicemail that pretends to return a call. Decline in one line and write the fair version.
  • Sections the user did not ask for.

When to use

Planning or debriefing a discovery, qualifying or demo call, trial or pilot; deciding what to say to a stalled buyer; answering an objection or a request for more material; writing a phone opening or voicemail; critiquing call notes or a demo plan.

When not to use

  • Written messages and sequences, including the same-day recap and any follow-up: email-and-sequences, if installed. This skill chooses the move; that one words it and sets the stop rule.
  • Whether a lever is fair, a trial converts fairly, or a line is pushy: fair-persuasion. Whether a number, reference or comparison is true: honest-claims. Terms, safety nets and the reply to "more than we expected": offers-and-value-propositions. What to charge: pricing-and-packaging.
  • A conversation whose purpose is to learn, not to sell: user-research-basics. Do not blend the two in one call. Slides for a pitch: presentation-design.
  • All if installed; if a neighbor is missing, use ordinary judgment.
  • Explicit user instructions, an existing sales process, a buyer's procurement steps and platform or legal rules (calling and recording consent vary by place) win over every default below.

If the plan already works

If a call plan, script or demo already has a goal, a gate, buyer-owned numbers and a dated next step, say so and change at most one thing. Do not add questions or stages for taste.

Facts first

List what the user supplied about the buyer and its source, what is unknown, and what you assume. Write [to learn] or a bracket for any missing figure and still deliver. Ask only if the chat is interactive and one answer would change the goal. Never fill a gap with a plausible number, quote or customer.

Decisions, in order

A small task starts at the decision it touches.

  1. Goal of the call. Default: one goal, written as the first line, from this list: learn whether the problem exists and is worth solving; qualify the account and the contact; book a named next meeting; show proof of abilities already agreed; reach a yes or no on a stated proposal; close out kindly. Change when the user names another single goal; two goals become one plus "next call".
  2. Verify, then ask. Default: confirm what is already known in one sentence ("I saw X; is that still right?"), then spend the questions on what is not. Change when nothing is known: say why you are asking and what the buyer gets from answering, then ask for the situation. See discovery and qualifying.
  3. Size the problem in the buyer's units. Default: each question names the number it should yield (how often, how long, how many people, what a miss cost last time, the target and why that target). A vague answer gets "how many, against what you need". No figure yet: [to learn], and getting it is the next step. Change when the stake is not money (safety, standing, a date): use the buyer's stated outcome and say it is not priced.
  4. Qualification gate. Default: before any demo, trial, pilot or proposal, three things are on the table: a problem someone feels and can size, a person who signs or a named route to a signature, and what the buyer will put in (time, data, a second stakeholder). If one is missing, the next step is to get it, not to show. Change when the purchase is small, repeat or self-serve: skip the gate and answer the question asked. See demos and trials.
  5. Fit, shrink or step out. Default: when the problem is small or the product does not touch its cause, say so on the call, shrink the offer to what matches, or point to a better fit, and give the reason. Change when the account is poor but the contact can introduce a good one: ask. Widen a small problem only by finding real impact, never by describing it larger.
  6. A stuck buyer. Default: a buyer who said they want to change and then goes quiet, says "I need to think" or asks for a fifth piece of material is afraid of choosing wrong, not unaware of the problem. Do not argue the cost of waiting again; give one recommendation with its reason and the condition that would change it, narrow what is left to review, and make the first step small and reversible. Change when the buyer never agreed the problem exists (back to discovery), named a dated step they are waiting on, or already has a clear, informed preference (confirm it and move to the next step). See stalls, objections and next steps.
  7. Objections and price. Default: tie the objection to the outcome the buyer stated and ask where it sits against that, rather than defending the product; price comes after value, as one figure with what it covers, then stop talking; a direct price question gets an answer or a range with what moves it. Change when the objection is a real gap in what is sold: say so and say what you would do.
  8. A dated, owned next step. Default: every plan ends with one next step naming who, what and when, said aloud, on both calendars, and confirmed in writing the same day. Change when the buyer says no: confirm it, record the reason in their words, stop. A phone opening or voicemail has its own shape: calls and voicemail.

Output

Deliver the plan, script, reply or critique first, with nothing before it. A call plan opens with its goal line and the next step it should end on, then the confirmed facts, the questions in order each tagged with the number it yields, and the gate check. A critique quotes each weak moment, says what it cost in a clause, and gives the replacement line, ranked: goal and fit, discovery, the showing, wording. Honor a stated count or length and count it yourself (a 30-second voicemail is about 75 spoken words). Add at most two one-line notes, only for what the user must fill in, verify or have checked.

Quick checks

  • The first line names one goal from the list, or the user's own single goal.
  • No fact the user supplied is asked as a question; every unsupplied figure is a bracket or [to learn].
  • Each question says which number it yields.
  • No demo, trial or proposal appears before the gate is met.
  • A stuck-buyer reply contains a recommendation with a flip condition and no fresh fear or deadline.
  • Price follows value; no concession answers a first reaction.
  • The plan ends on a next step with who, what and when.
  • Nothing in it pushes after a no or hides a term.

References

  • discovery-and-qualifying.md: load for a discovery or qualifying plan, a question list, call notes to review, a kind decline, or a second stakeholder.
  • demos-and-trials.md: load for a demo plan or critique, handling questions in a demo, or any trial, pilot or "can we try it first" request.
  • stalls-objections-and-next-steps.md: load for a silent or stuck buyer, an objection, "send me more information", a price pushback, a lost deal or the close of any conversation.
  • buying-groups.md: load when several people must agree (signers, a security or finance review, "I'll take it to the team") or a deal stalls after one contact's yes.
  • calls-and-voicemail.md: load for a cold-call opening, a voicemail, a gatekeeper or booking a meeting by phone.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.